MFL-Based Local Damage Diagnosis and SVM-Based Damage Type Classification for Wire Rope NDE.

Area of Science:

• Materials Science
• Non-destructive Testing
• Mechanical Engineering

Background:

• Wire ropes are critical components in elevators and cranes, essential for lifting heavy loads.
• External environmental factors can cause wire rope damage, including breakage and cross-sectional loss.
• Damage to wire ropes under tensile stress can rapidly expand, posing significant safety risks due to stress concentration.

Purpose of the Study:

• To apply the magnetic flux leakage (MFL) method for diagnosing various types of damage in wire ropes.
• To analyze the characteristics of MFL signals corresponding to different damage types.
• To develop an automated system for classifying wire rope damage.

Main Methods:

• Damaged wire rope specimens were scanned using a multi-channel sensor head and a compact data acquisition system to measure magnetic flux leakage signals.
• Signal processing techniques, including a Hilbert transform-based enveloping process, were employed to reduce noise and enhance signal resolution.
• A support vector machine (SVM)-based classifier was trained using extracted damage indices for automated damage classification.

Main Results:

• The study successfully verified the possibility of diagnosing several types of wire rope damage using enveloped magnetic flux signals.
• Distinct characteristics of MFL signals were identified for different damage types through the analysis of extracted damage indices.
• The trained SVM classifier achieved automatic classification of damage types, distinguishing between cutting and other forms of damage.

Conclusions:

• The magnetic flux leakage (MFL) method is effective for diagnosing wire rope damage, including cuts and corrosion.
• Signal processing techniques significantly improve the resolution and interpretability of MFL signals.
• Automated classification of wire rope damage using SVM provides a reliable method for ensuring structural safety.